How Do I Generate a Knowledge Graph from a News Feed, using Virtuoso?

kidehen · July 22, 2021, 4:39pm

I recently stumbled upon a Blog Neo4j Post about making sense of news via a Knowledge Graph generated from content published via the [CNET News Site] (http://news.com/).

Naturally, I decided to replicate the same process using our Virtuoso platform with the generation of a Knowledge Graph deployed using Linked Data principles as my goal.

Courtesy of Virtuoso’s built-in data access and integration middleware functionality (provided by the Virtuoso Sponger Module) I completed the task in two simple steps.

Locate RSS feed — OSDS simplifies this via its Plain Old Semantic HTML (POST) tab which makes the xhv:alterante property (relation) easier to locate, if it exists
Once found, construct a Virtuoso Sponger URL that incorporates CNET’s RSS Feed URL

That’s it, you have a clickable Knowledge Graph that helps you make sense of the news from CNET’s News Portal.

Alternatively, I can achieve the same results using OpenLink Structured Data Sniffer (OSDS) Browser Extension via the following steps.

Goto CNET News Home Page
Initialize OSDS by clicking on its icon
Goto the Plain Old Semantic HTML (POSH) Tab to view document metadata
Located the CNET RSS Feed URL
Select Download to SPARQL Endpoint Option
Execute

The end result is a SPARQL Query Results page that includes a schema:CreativeWork entity identified by a Hyperlink. Click to retrieve an HTML document comprising its full description from the newly generated Knowledge Graph – presented in property-sheet form, and then take the xhv:alternate → schema:mainEntity → property path to a schema:DataFeed instance comprising each news item.

Courtesy of buit-in Faceted Browsing functionality, I can also explore new items from other sources across the evolving Knowledge Graph collective maintained by our public URIBurner instance.

Here’s a screenshot sequence that covers actions performed in this exercise.

Metadata Discovery